ARCTOS ACT A PHILOLOGICA FENNICA

VOL. XXX

HELSINKI 1996 HELSINGFORS

INDEX

ROLF WESTMAN History of Classical Scholarship in Finland: 7 A Bibliography

PATRICK BRUUN Two Facets of Ancient Monetary Economy: 21 Celtic Imitations and Roman Rigid Formality

ANNE HELTTULA Truffles in Ancient Greece and Rome 33

MAARIT KAIMIO How to Enjoy a Greek Novel: 49 Chariton Guiding his Audience

MIKAKAJAVA New Poems on Stone 75

KALLE KORHONEN On the Composition 101 of the Hermeneumata Language Manuals

MARITI LEIWO Language Attitude and Patriotism. 121 Cases from Greek History

JARI PAKKANEN The Height and Reconstructions of the Interior 139 Corinthian Columns in Greek Classical Buildings

OLLI SALOMIES Observations on Some Names of Sailors 167 Serving in the Fleets at Misenum and Ravenna

RAIJA SARASTI-WILENIUS Do tibi me totam. Latin Wedding Poetry in Finland 187

MARY SIANI-DAVIES Gaius Rabirius Postumus: 207 A Roman Financier and Caesar's Political Ally

HEIKKI SOLIN Analecta Epigraphica CLXIV-CLXVI 241

RAIJA V AINIO A Reading in Consentius Reconsidered. 247 A Case of Palatalization

ROLF WESTMAN Possible One-verse Additions 257 before Eur. Supplices 263

De navis libris iudicia Index librorum in hoc volumine recensorum Libri nobis missi

261 313 317

ON THE COMPOSITION OF THE HERMENEUMATA LANGUAGE MANUALS*

KALLE KORHONEN

1. Introduction

The bilingual Greek-Latin language teaching material, known as Her­meneumata or Interpretamenta Pseudodositheana, is a peculiar chapter in the history of ancient erudition. Published in the third volume of the Corpus glossariorum Latinorum (= CGL) in 1892, they have received little attention in the scientific literature of this century.1 Still, they give us plenty of infor­mation about language learning and teaching in antiquity, about ancient everyday life, and about the study of Greek in the Medieval West. 2 In this article, my aim is to discuss some central problems of this mat~rial: why do the texts seem to have two target groups; what do the three books mentioned

*I am grateful to Carlotta Dionisotti, Maarit Kaimio, Kaspar Kolk, Martti Leiwo, Roger Wright and many other friends and colleagues for comments and help.

1 For the essential discussion see: G. Goetz, CGL III (1892) vii-xxxi, xxxiv-xxxvi; Id., CGL I (1923) 12-23, 284; A.C. Dionisotti, "From Ausonius' Schooldays? A Schoolbook and Its Relatives", JRS 72 ( 1982) 83-125 [ = Dionisotti, Schoolbook], especially 86-94. The article contains the edition of a previously unknown Hermeneumata text. -For an overview, see H.-I. Marrou, Histoire de !'education dans l'Antiquite7 (= Marrou, Histoire) II (1975) 59 and 193-94 note 20; J. Debut, "Les Hermeneumata Pseudodositheana. Une methode d'apprentissage des langues pour grands debutants", Koinonia 8 (1984) 61-85; A.C. Dionisotti, in: The Oxford Classical Dictionary3 ( 1996) 690. I will call the material simply Hermeneumata because there is no reason to connect them with the grammarian Dositheus, see H. Keil, Grammatici Latini VII (1880) 369 ff. and G. Flammini, "Prolegomeni all a recensio plenior degli 'Hermeneumata Pseudodosi theana"', GIF 42 (1990) 3-43 (= Flammini, Prolegomeni), esp. 3-5.

2 On the last subject, see especially B. Bischoff, Mittelalterliche Studien II (1967) 260-61; A. C. Dionisotti, "Greek Grammars and Dictionaries in Carolingian Europe", in The Sacred Nectar of the Greeks: The Study of Greek in the West in the Early Middle Ages (ed. by M. W. Herren) (1988) 1-56 (= Dionisotti, Grammars & Dictionaries), esp. 26-31, and B.M. Kaczynski, Greek in the Carolingian Age. The St. Gall Manuscripts (1988).

102 Kalle Korhonen

in the prefaces contain; and finally, problems of the dating and the original context of the collections.

I will restate briefly the basic facts of the Hermeneumata. More or less complete collections have survived in ea. 50 Western manuscripts and early printed books. The material contains the following bilingual parts:

a) words, mostly verbs, in alphabetical order (Greek-Latin) b) noun lists arranged by topic (KE<paAata- capitula) (Greek-Latin) c) texts for language practice.

In the collections, the most characteristic parts of the text material often have the title Ticpt Ka9ru.tcptvfls OJ..UAtas or cruvavacrtpo<pfls - De sermone cotidiano or de conversatione cotidiana. These texts have not survived else­where, and they may contain the description of a single day from dawn to dusk. The whole description consists of scenes: morning, school (and pos­sibly lunch), official and social routines, bathing, dinner and the prepara­tions for night. The texts are traditionally called colloquia, and I will use this plural term in this article, even though it has some disadvantages. In some versions other texts are included, such as fables of Aesop, rudiments of Roman law, mythology etc.

The editor of the CGL, G. Goetz, suggested that the parts originally formed a language teaching material in 12 books, and that one surviving version (Hermeneumata Leidensia) was close to the original.3 The first pro­position has been convincingly refuted, because 1) it is unlikely that all the 12 texts put together by Goetz belonged to the same collection, 2) the men­tion of 12 books, which appears only once, is clearly an addition by a later systematizer, and 3) the informative, though somewhat chaotic, prefaces of the texts emphasize that the collection is in three books. 4 Because of the consular date of the year 207 in his "original" version, Goetz believed that the manual was created at this time, and imitated by various schoolmasters throughout the remainder of antiquity.5 The description of the 207 version as being the original one has been strongly criticized by A.C. Dionisotti.6

3·G. Goetz, CGL I (1923) 17-19 following on the lines of H. Keil, Grammatici Latini Vll (1880) 374, note.

4 Dionisotti, Schoolbook 90 (supported by Flammini, Prolegomeni 42 n. 83). The num­ber of books: see CGL Ill 7,70-8,15; 30,21-42; 119,24-26; 120,5-6; 166,10-19; 283,15-16; 289,22. About the four-book collections see below, n. 41 p. 110.

5 G. Goetz, RE VII (1912) 1438; CGL I (1923) 17-19. He uses the term enchiridion.

6 Dionisotti, Schoolbook 86-92.

On the Composition of the Henneneumata Language Manuals 103

The Hermeneumata, in their current form, are not a single manual. On the basis of the manuscript tradition, the versions can be grouped as follows. Their contents are also listed (ABC =alphabetical glossary, cap. =capitula glossary, coll. = colloquia, texts = other texts). I also list the conventional abbreviations of the versions.

B = Henneneumata Bruxellensia (CGL lli 393-98, 398-421) ABC, cap.

C = Henneneumata Celtis (Dionisotti, Schoolbook 97-106 + Wien, ONB Suppl. gr. 43, ff. 18- 45v)

cap., coll. E = Hermeneumata Einsidlensia (CGL Ill 221-79)

cap., coll. H = Hygini Henneneumata (CGL Ill 72-94)

ABC, cap. L = Henneneumata Leidensia (CGL lli 1-72 [coll. 637-38]; 94-108)

ABC, cap., coil., texts M= Hermeneumata Monacensia (CGL lli 117-220 [coll. 644-54]) 7

ABC, cap., coil. Mp = Hermeneumata Montepessulana (281-343 [coll. 654-59]; 487-506; 506-31)

ABC, cap., coil. S = Henneneumata Stephani (CGL lli 345-90, 438-74, 474-87);

ABC, cap., 2 coil. (S 1 = coll. I, 376-79,8 S2 = coll. ll, 379-84) V= Henneneumata Vaticana (CGL lli 421-38).

cap.

The textual history of the different versions is quite complicated, 9 e. g., the relations between Hermeneumata Leidensia, Hygini H., and H. Stephani are

7 M and E are so close to each other that they clearly have a common origin. E, in which the Greek text is written in the Greek alphabet, is not based on M, which has only the Latin alphabet. See Goetz, CGL Ill, xxii. The M and E colloquia will be quoted together.

8·The beginning of the S1 colloquia is similar to the one in L, but S1 continues after L finishes. I will cite these colloquia together.

9 It has been studied by K. Krumbacher (De codicibus quibus Interpretamenta Pseudo­dositheana nobis tradita sunt, Diss. Munich 1883 and RhM 39 [1884] 348-58), G. Goetz [seen. 1], and A.C. Dionisotti (Schoolbook 86-90; "From Stephanus to Du Cange: Glos­sary Stories", RHT 14-15 [ 1984-85] 303-36 [= Dionisotti, Glossary Stories]; Grammars & Dictionaries 26-31). As to the L version, see also Flammini, Prolegomeni 9-43. For a list of the versions and manuscripts see Dionisotti, Schoolbook 87 and Grammars & Dictionaries 27-28.

104 Kalle Korhonen

extremely difficult. The position of the Colloquium Harleianum (CGL Ill 108-116 = 638-44), which belongs to the Hermeneumata material on the basis of its title and contents, is unclear. I will cite it separately as Coll. Harl. 10 I would also point out that the order of the three basic elements is variable in the manuscripts. Two papyrus fragments of colloquia have sur­vived, P. Prag. II 118 (5th c.),11 and P. Berol. inv. 10582 (5th or 6th c.).12 The first one contains fragments of the Coli. Harl. (6; 8-9 = 109,73-110,2; 11 0,29-42); the second has two colloquia which do not directly belong to any of the manuscript versions, but have some similarities with Mp.13

Before proceeding, some points must be made about the language of the Hermeneumata. The language in the colloquia is very close to spoken language.14 There is no tendency to teach "correct" forms of language in the manner of grammarians.l5 On the other hand, the syntax and vocabulary, especially in the Greek part, show a strong interference of the other lan­guage.16 "Spoken language" always has a certain context, but in this case finding out when and where this kind of language was taught is extremely difficult because first one should make a clear distinction between the features of spoken language and those produced by interference.17 Another problem is that since all the surviving manuscripts are western, the Greek

10 The L, H, M, and Mp colloquia, edited by Goetz in CGL Ill 637-59, will be quoted by his chapters, all the other parts of CGL Ill with page and line numbers, and C colloquium by the chapter numbers given by Dionisotti. The letters (except C) thus always refer to the version in CGL Ill; the single manuscripts will be indicated when quoted exclusively.

11 Ed. pr. J. Kramer, in: Papyri Graecae Wessely-Pragenses 11 (1995) 3-5. 12 W. Schubart, Klio 13 (1913) 27-33; CPL 281; J. Kramer, Glossaria bilinguia in papyris et membranis reperta (1983) (= Kramer, Glossaria) 99-103 n. 15. 13 The bilingual papyrus PSI VII 848 = CPL 39 contains L 45,42-48 and 46,9-17 (fables of Aesop).

14 Cf. L. Zgusta, in: Die Sprachen im romischen Reich der Kaiserzeit (1980) 124-25. 15 In fact, the author of the Appendix Probi was using Hermeneumata noun lists to find unsuitable and incorrectly written words, as C.A. Robson has shown (MA 69 [1963] 37-54). To his examples may be added four words belonging to the category de cognatione (Prob. app. gramm. IV 198,35 - 199,1); two or three of them are in M (nurus 181,57; socra 181,55; ancula (?) 181,37). However, the "incorrect" forms in Appendix Probi usually deviate much from the forms in the Hermeneumata manuscripts printed in CGL.

16 Cf. Dionisotti, Schoolbook 92. 17 There is one approach which could also be used to analyse the texts, namely that of pragmatics (a field of modem linguistics), because we know little about ordinary conver­sations in antiquity. I am planning a study on the linguistic aspects of the texts.

On the Composition of the Hermeneumata Language Manuals 105

texts have sometimes suffered badly; in M, the Greek has been written in the

Latin alphabet. However, it is important not to date them a priori to Late

Antiquity. I will cite both the Greek and the Latin versions without any

syntactical emendations (except when the C version is concerned).

2. A phrase book as a schoolbook

When looking at the contents of the Tiept Ka8nJ..LEptvfl\; OfltAta\; -De sermone cotidiano parts, 18 one has the impression that the texts have not

originally had a single target group, but two: children and adults. Compare,

e.g., the two passages from M: llpoflA.Sov EK 'tOU KOt'tOOVO\; cruv 'tql 1tatba­

yroyfp Kat cruv ttil ttpocpfp acrnacracr8at 'tOV 1tattepa Kat 'tllV J.lll'tEpa- Pro­cessi de cubiculo cum paedagogo et cum nutrice salutare patrem et matrem

(M 2), and KuptE, tti £ntttacrcret<;; Mflttt exet<; xPilJ.latta euKatpouvtta; T{

XPEtav EXEt\; 8ave{cracr8at; Ei EXEt\;, xpflcr6v J.!Ot 1tEV'tE Bnvapta. - Domi­ne, quid imperasti? Numquid habes pecunia< m> vacua< m>? Quid opus habes mutuari? Si habes, commoda mihi quinque sestertia{s} (M 5 = 212,46-54), or from C: Aeaivro Kat napaypa<pro npo\; ttov £niypaJ..LJ..LOV ... Kat bEtKVuro

'tql btbacrKOV'tt J..LE. Kat E1tatV1l<JEV flE O'tt KaAOO\; eypa'tfa. - Deleo et prae­duco ad superpostum . . . et ostendo doctori meo. Et laudavit me quod bene scripsi. (C 27), and Tt\; outtro<; 1tOtEt 00\; cru, tva 'tO<JOU'tOV 1ttU\;; T{ EipftKacrt

oi 18ovtte<; cre ttotouttov ... Toutto 8£ npenEt cpp6vtJ..Lov oiKobecrn6tt11V iBto­

npayf..Lova ... £auttov eu8uvetv; - Quis sic facit dominus quomodo tu, in tan­tum bib is? Quid dicent qui te vide runt talem ... Ita hoc decet sapient em pa­trem familias sui negotii ... semet ipsum regere? (C 66). In the colloquia, the

protagonist in the morning and at school is a schoolboy.19 The persons who

participate in private and public affairs are adults, which is also the case in

the dinner descriptions and in the bathing sequence.20 Another feature that

seems to separate these parts from each other is the manner of narration. In

the schoolbook parts, the schoolboy acts as the narrator, he keeps telling

what he does as the story proceeds; naturally, there is also some dialogue. In

the phrase book there is no such narrator, only dialogue. In fact, the name

18 For an overview of the contents, see also Dionisotti, Schoolbook 93-94.

19 The only morning scene with an adult man is in Mp 4.

20 The difference in the protagonists has also been noted by Dionisotti, Schoolbook 93-94, but she prefers to divide the material into scenes.

106 Kalle Korhonen

colloquia better suits these sequences. If we go through the versions one by one, we can distinguish the two

types of material and see if all the colloquia conform to these rules. The L/S 1 colloquia cause problems. L finishes in eh. 8, in which the boy arrives at the baths. S 1 which, as I noted above, is a longer version of the same text, continues with bathing, oaths and dinner (mostly names of foods and types of wine) (378,32- 379,66), and the age of the protagonist is not clear. The first-person narration continues through the colloquia. Anyway, it seems to show some editorial effort (by the publisher Estienne?) to even out the dif­ferences between the two distinct parts. Coli. Harl. has both the schoolbook (3-10) and the phrase book elements (11-28). It also has a preface, a dia­logue between the boy and his father (1-2). The schoolbook part is followed by a title which well suits a phrase book.21 Here the problem is that it has, in the school part, mostly dialogue. The only exception is c. 9, in which events of adult life are described, but in third-person narration. Fortunately, the chapters 8 and 9 can be connected with the aid of the papyrus P. Prag. II 118, the last sentence of which supplements the London ms. of Coli. Harl.: ~pEv yap f.l~ o na'tnp f.lO'? Ei~ 't? [n]pat'tro[ptov] J..LE9' £au'tou (11. 17-20).22 This short description is clearly a part of the school text, but as I said, Coli. Harl. has mostly dialogue.

The order of M is interesting: the schoolbook sequence (2) is followed in the manuscripts by the two glossaries, and then come the beginning of the same schoolbook sequence (3) and a phrase book part ( 4-12). This time, the type of narration is first-person and dialogue, respectively.23 Mp has a pre­face, a dialogue of a pupil and his teacher (Mp 2) which is followed by a dialogue sequence with adults (3-20). S2 only has the schoolbook sequence (379-384,29); it begins with: Lege bene. Hodie quid fecisti?- 'Avayvco9t KaAffi~. Lllf.lEpov 'tt E1tOtllO"a~;- (379,68-70), and continues with first-person

21 H 11: TiaAtV ep& ()UJ!J.lLK'ta Kat avayKata. "Ecr'ttV 8£ -rau-ra acrnacrJ.lO~ A.6yrov'

enepro-rilcret~, A.otOopiat Kat &A.A.a noA.A.a. - Iterum dicam commixta et necessaria. Est aut em haec salutatio sermonum, int'errogationes, maledicta et alia multa.

22 It must be noted that excerpts from Coli. Harl. can also be found in Leiden, UB Voss. lat. 24, Bern, ms. 236 (see Goetz, CGL Ill xxxi; CGL I 22-23), and in Paris, BN lat. 7683, of which at least the Paris ms. is an independent witness (see Dionisotti, Glossary Stories 329-30). 23 In M 4, the phrase book sequence begins with some third-person narration, which ends after few lines.

On the Composition of the Henneneumata Language Manuals 107

narration. C, which, as Dionisotti says, 24 has been put together from two or more different sources, has both parts (1-46; 47-69); the narration follows the usual manner. The titled sequence 70-77 is like an expanded version of Coli. Harl. 9. The chapters have a didactic tone, and the main character is the father of the person to whom the story is told (C 71). I believe that the sequence belongs to the schoolbook part. The Berlin papyrus shows no tra­ces of a schoolbook part, but has, in the middle, a title crEeJ:Lco Kco[ ttoux.]­vouc;- oJ.tf t]A~~ Ka811J.1Eptvn (11. 42-43). Therefore, it probably is part of a larger phrase book which has contained many colloquia.25

To conclude, it seems that it is useful to distinguish two different kinds of material in the colloquia. I will call these different parts "school­book" and "phrase book". The schoolbook parts have to do with the early stages of literary education, and they give us much information about an­cient schools. Such narrations may have been used as translation practice. (I hope to analyze the school parts further in another study.) The literary genre of phrase books has not been studied extensively. To use a definition pro­posed by N. Haastrup, they represent "idealized dialogues that are meant to be used as models for verbal conversation in specific situations". 26 It seems clear that these parts are meant to be an aid in the acquisition of a foreign colloquial language. They would not exist in unilingual form. They also de­monstrate in which situations to use the expressions, as the title I1 E p t Ka811J..LEptvf1<; OJ.ltA{a<; -De sermone cotidiano indicates.

It is, of course, possible to say that even though the protagonist changes, the material may still have been written for children in order to teach them both language and everyday affairs. After all, there are dialogues in grammars, too. It is true that the surviving versions seem to have been used by schoolchildren, as they have been transmitted among grammatical material. But writing bathing and dinner descriptions that are meant to be used at school before the children start with Homer and Vergil does not seem reasonable. What if the material has been written for both adults and children because foreign language acquisition was necessary for both? Should we rather divide the material only into scenes- traces of such a divi­sion have survived in C, 27 or into different colloquia or fabulae cottidia-

24 Dionisotti, Schoolbook 94, 120.

25 Cf. W. Schubart, Klio 13 (1913) 34.

26 N. Haastrup, in: Symposium on Lexicography Ill ( 1988) 390.

27 C, title: Ke<paAea VOV'f\1tept Ka9tjleptV'f\<C> avacr'tpO<p'f\c; C 70: de lucubris et negotiis

108 Kalle Korhonen

nae?28 In my opinion, however, the clear differences in the contents of the

texts show that the division proposed here is useful.

In CGL I, G. Goetz made an important distinction, when he divided

the early bilingual glossaries (glossaria vetustiora) in two groups: those

based on the studies of the Roman grammarians, and those on the communi­

cation of Romans and Greeks. To the first group belong the etymologies

(etyma), significations and idiomata; to the second hermeneumata, which, as

I said earlier, Goetz saw as a manual created in the third century.29 He noted

that the glossaries are only loosely, if at all, related to the other extant

Greek-Latin or Latin-Greek vocabularies. He added that they had been com­

posed using material collected by lexicographers; 30 it is true that the ar­

rangement of the capitula titles belongs to the 6voJ..LUO"'ttK6v tradition.31 I

think we can agree with Goetz here and see the Hermeneumata mostly as

something that does not belong to the grammatical tradition. But what about

the school parts? After all, they contain explicit references to the study of

grammar, grammatical terms, etc. (M 2; S2 381,28 - 382,73; C 18-42).

There are also, in the capitula glossaries, sections called de studiis (L 24-25;

S 351; title in H 82,40), de ludo litterario (M 198-199 [E 277-278]; Mp 327;

S 351-352; Cf. 34r-34v) or de instructione artis grammaticae (Mp 327-328; s 375-376).32

In all, it seems that the Hermeneumata contain at least two kinds of

originally separate material. The parts must have been put together already

in antiquity because most of the extant versions now have both elements, the

only exceptions being S2 colloquia and the Berlin papyrus. 33 All the

forensibus. See Dionisotti, Schoolbook 94.

28 The discontinuity between the scenes was noticed by K. Krumbacher, in: Abhand­lungen aus dem Gebiet der Klassischen Altertums-Wissenschaft. W. von Christ zum 60. Geburtstag (1891) 309. He suggested that the different colloquia be numbered separate­ly, a principle that Goetz partly followed in CGL Ill 637-659, although he used names like "Colloquium Leidense" etc.

29 CGL I 13: Aut e studiis succreverunt grammaticorum Romanorum, qui ... latina voca­bula cum graecis diligenter contulerunt, aut e commercia Romanorum et Graecorum, quod Romanos litteras graecas, Graecos latinas discere coegit. Cf. op. cit., 13-22, 284. 30 Goetz, CGL I 22.

31 C. Wendel, RE XVIII (1939) 515-16. On the capitula glossaries, see also I. Schoene­mann, De lexicographis antiquis qui rerum ordinem secuti sunt, Diss. Hannover 1886.

32 C also has flept ayroyi1<9 Kat a~qn~aaero<; (f. 31v-32r).

33 The text of the recently published P. Prag. II 118 entirely belongs to the school part as

On the Composition of the Hermeneumata Language Manuals 109

versions we have were used as schooltexts, which is shown also by the re­ferences to Hermeneumata in the school scenes.34 This has granted the sur­vival of the adult material. The distinction hermeneumata vs. material based on grammarians' work is useful, but needs to be applied carefully. The pre­faces of the texts can tell us more about the original purposes of the different parts.

3. The three books and their prefaces

As I stated before, the repeated references to three books in the pre­faces make it clear that at one stage this was the form of the material. I think those of the books I and II indicate that three-book collections existed for practical purposes.

The prefaces of the alphabetical glossary (the first book) and the capi­tula glossary (the second book) are quite similar in the different versions, and they all seem to be based on the same model. The preface of the first book, which also is the preface of the whole collection, survives in its origi­nal place in the London manuscript BL Harl. 5642,35 and in two manu­scripts of the B group.36 We have it also in M 137 and in Mp 1. In the two

defined by me.

34 M 2: EKJ.lav8avro EPJlT\VEUJla'ta- edisco interpretamenta; C 34; compare S 2 381,59-60: eypawa KaenJleptva - scripsi cotidiana. There is one rather unknown reference to hermeneumata in ancient literature which may be relevant here. Seneca the Elder seems to talk about this kind of material in contr. 9,3,14 (cf. 9,3 exc.). He mentions an orator who gave speeches in both Latin and Greek, and when some people complained that he was being paid too little, Q. Haterius Agrippa (cos. suff. 5 BC) said: numquam magnas mercedes accepisse eos, qui hermeneumata docerent. This immediately brings to mind the poorly-paid primary teachers. However, the significance of the word hermeneumata remains unclear; we are dealing with a very early period, and it may refer to translation practices mentioned by Suetonius (gramm. 4 ).

35 Goetz prints the preface in CGL Ill ix. It is not so clear as the prefaces in M and Mp, and has elements not found in B, M, or Mp. The same ms. also contains the Coll. Harl. SeeK. Krumbacher, RhM 39 (1884) 348-50. 36 They are Leiden, UB Voss. lat. F 26 (the so called Glossarium Leidense, CGL Ill 398-421) and Angers, mss. 477, omitted by Goetz and published by H. Omont, BECh 59 (1898) 671-88; the preface: 675. The composition of the Leiden manuscript has been dis­entangled by Dionisotti, Glossary Stories [see n. 9] 305-12. In it, the lines of a Herme­neumata collection have been turned from Greek-Latin to Latin-Greek and arranged

110 Kalle Korhonen

last cases, it is no longer the preface of the alphabetical glossary, 38 but the

preface of the colloquia. I quote the Mp version here:

'EnEtO'h op& 1tOAAOU~ E1tt8UJ.!OUV'ta~ (EAA11Vt<J'tt OtaA.£yecr8at Kat

(ProJ..tatcr'tt Jl~'tE euxep&<; 8uvacr8at ... ouK e<pEtcraJ..tllV 'tou'to norilcrat

Yva ev 'tptcrtv ~t~A.iot<; EPJ.!llVEUJla'ttKot<; nav'ta 'ta p~J.!a'ta I

cruyypa'JIOOJlat. Quoniam video multos cupientes Graece disputare et Latine ne que fa­cile posse ... non peperci hoc face re ut in tribus libris interpretatoriis omnia verba conscribam.

The first book will contain the letters from A to Q. The preface in M (and E) is longer than in the other versions, but the essential contents are the same.

Some more information about the contents of book I can be obtained from

the prefaces of the second. 39 In them, the author states having given p~Jla'ta

Kat 'tOU'tOOV EK J.!Epouc; avayKata El<; KAt<JtV Pllflcl'tOOV - verba et eorum ex parte necessaria in declinatione verborum (L, H)40 in the first book.

According to the prefaces of book II (seen. 39), the author will write

in it nept nav'trov (A.otn&v Mp) npayJ..ta'trov Kat 'ta Ke<paAata au'tmv - de omnibus rebus et capitula eorum (M, cf. Mp and L 30,55- 31,2). Some of

these prefaces also contain a table of contents of the capitula (H 82; M 166-

167 [E 235-236]; C 18r-18v [p. 92-93 Dionisotti]). In the prefaces of books

alphabetically. The list of words belonging to the preface can be found in op. cit. 307.

37 E has basically the same preface (223 - 224,39), but the order of the books has been modified in it.

38 M has an addition which joins the preface to the colloquium (the last ten lines in M 1 = 120, 17-37). In the other versions, the alphabetical glossary begins with some introductory lines (L 3,26-29: a1toOrocrro oi>v 1:a A.ot1ta Ka"Ca cr7:otxe!ov; Mp 337,7,9: "Ap~acr9e a1to apxil~, cfr. H 72,1, M 122,60-61).

39 L 7,65- 8,19 and, strangely, 30,55-31,2 (preface of the "third" book), cf. also 47,58-48,7; H 81 - 82,7; M 166,10-29; Mp 289,21-43; fragments in Glossarium Leidense: 402,81, 405,76-78,407,42,408,44-45,415,64-65,418,6 (cf. Dionisotti, Glossary Stories [see n. 9] 307 n. 6).

40 The pftJ..la7:a- verba really seem to mean here "verbs", as the glossaries mainly contain verbs in their different forms. The not so numerous words from other parts of speech have only one form. The verba quae pertinent ad artem grammaticam in the confused preface of the ms. Harl. 5642 (CGL Ill ix) is not consistent with the contents of the ABC glossaries in general, but seems to be an addition, cf. also the contents of the same ms. as listed in CGL Ill ix.

On the Composition of the Hermeneumata Language Manuals 111

I and 11, the authors do not specify for whom the bilingual word-lists have been written; the words used are av8pronot- homines (L 8,7-10 [book 11]:

OU'tro<; EUKOAffi<; 'tll<; OJ.ttAta<; -r&v av8pronrov EUXPT\O'ta EO''t<Xt - uti facilius sermoni hominum proderit, cf. H), or 1tOAAo{ - multi (M + E, Mp, book 1). This means, in my opinion, that the original target group has been adults rather than children. It is remarkable that the languages mentioned in these prefaces are always both Greek and Latin; the alphabetical and capitula glossaries have been far n1ore useful for Greeks who have wanted to use Latin.

Could it then be that in an "original" form, a Hermeneumata collec­tion would have contained three books with an alphabetical glossary in the first, and the capitula glossary divided into the two last books? After all, the prefaces use the expression "all the words", so where do the texts come in? Such is the arrangement of the three books in the fragmentary B version. In this group of manuscripts we have explicits in the middle of a capitula glos­sary. 41 Otherwise, they always come in the end. 42 Therefore, it seems that in a typical version book 11 was sufficient for the capitula glossary.43 Anal­ternative solution would be that book Ill contained an alphabetical Latin­Greek glossary, in which case the three-book collection would really have been useful to the students of both languages, as the prefaces promise. But then we should have to conclude that these dictionaries would have dis­appeared without leaving a trace in the surviving manuscripts, which seems unlikely to me.

At this point, it seems safe to suppose that the three-book collections to which the prefaces refer had texts in the mysterious third book. But what

41 Dionisotti, Glossary Stories 307.n. 1. Book Ill in Glossarium Leidense 409,72; three books in the capitula glossary of the Angers ms. 477 published by H. Omont, BECh 59 (1898), 679, 682, 685. B 395,63-65 may indicate two glossaries arranged in four books. However, one ms. of the B group also had the S 1 colloquia, see Dionisotti, op. cit. 315.­The four-book Hermeneumata Vaticana is the product of a later reworking (on it, see I. David, in: Commentationes philologae Ienenses 5 [1894] 199-202; L. Traube, ByzZ 3 [ 1894] 604-06; G. Baesecke, Der Vocabularius Sti. Galli in der angelsachsischen Mission [1933] 80-81; B. Bischoff, Mittelalterliche Studien II [1967] 267).

42 Book I: L 7,61-63; H 81,50; M 166,9; book II: M 210,43 (possibly also the sign with an uncertain meaning in Mp 337 after line 6, see Goetz in app. crit.).

43 The extent of the glossaries in the manuscripts does not help us to solve the question, as the longest glossaries in CGL Ill are in M, both a little more than 40 CGL pages long (122,62- 166,8 and 166,30- 210,43). The prose parts are in no way as extensive.

112 Kalle Korhonen

texts? Can we connect the extant prefaces of book Ill with those of books I and II? The only preface of book Ill that mentions the preceding books is the one in L (30, 14 - 31 ,23). It is a strange mixture. Its first part seems to be written for a single person, possibly an adult - because the children of the student are mentioned - who "loves to speak Latin". 44 The contents are de­fined only as "something with which to practise". Then follows a fragment of the preface of book II (30,55 - 31 ,2).45 It seems that the author has at first thought to list here the titles of the texts to follow, and used an expression from the preface of book II. However, we here have only the prologue of the text Hadriani sententiae (31,3-23). It is completely uncertain whether the first part belongs to the series of the first two prefaces or to the collection that follows.

In the preface of the colloquia in M (M 1 = 120,17-39, cfr. E 224,18-39), the author specifies the purpose of the book: E1t£t8~ vrpttOt<; natatv ap­

XOJ.lEVOt<; 1tatb£U£a8at avayKa'iov ll.f>pov aKp6aatv EPJ.lllVEUJ.latrov OJ.llAt­a<; Ka811J.1Eptvfl<; - quoniam parvulis pueris incipienti<bu>s erudiri necessa­rium videbam auditionem interpretamentorum sermonis cottidiani. This will help children to learn to speak Latin and Greek. In Mp, the preface of the colloquia (Mp 2) is a teacher-pupil dialogue; Latin seems to be the more foreign language. Coll. Harl. has been used by Greek -speaking school­children (1; 3-4). These prefaces, both of which follow the preface which actually belongs to the first book, clearly reflect the school use of the material.

The prose parts in the other H ermeneumata versions have neither pre­faces nor book Ill incipits (in spite of the book II explicits), only the title Ocpt OJ.ltAta<; Ka811J.lEptvfl<; -De sermone cotidiano, or something simi­lar.46 This is even mentioned in the table of contents of M capitula

44 I need to cite it at length: (in the two books) auv£ypawa n:av'ta 'tU PT\I.la'ta, a i18uv~-8'llv 'tTI 'h!lc'tEpq; EP!l'llVctq;, oaa avayKata U1tOAa!l~avro Kat oAro~ oaa cO<p£A£t av8pro-

1t0t~ <ptAll'tat~ 'tll~ AaAia~ pro/latK:ll~. OUK £8ia'taaa Kat £v 'to{nql 'tip ~t~Atqln:poa8£t­vat, tva £xn~ on:ro~ £au'tOV YU!lVaan~, cXAAa Kat £U'tuxro~ 'tEKVOt~ <JOt~ Ka'taAtn:n~ !lV'll­!lO<JUVOV Kat un:60ctY!la qnA07tOVtrov a&v - conscripsi omnia verba, quae potui nostra interpretatione, quae necessaria arbitror et omnino quae prosunt omnibus amatoribus loquellae Latinae. Non dubitavi et in hoc libro adicere, ut habeas, ubi te ipsum exerceas, sed et feliciter liberis tuis relinquas memoriam et exemplum studio rum tuorum (30,23-48). This is the only place in the prefaces in which only Latin is mentioned.

45 It is very similar to M 166,24-27 and Mp 289,32-36. 46 L, title= 69,41-43; Coli. Harl., title= 108,1-2; M, title p. 647 = 210,44-45; S1 376,47;

On the Composition of the Hermeneumata Language Manuals 113

( 167,24 ), where it comes last, and C, 47 where it is in the middle of the index (it may well have been book Ill even if it is in the table of contents of book 11). At this point we can conclude that book Ill has in some Hermeneumata versions begun simply with the title. It is better to discuss the L version se­parately. It also takes us to more general problems of context.

4. Problems of dating and context

The L version has an intriguing consular dating which has been used to date all the Hermeneumata. 48 In all, L contains, after the preface that starts book Ill (see above), a collection of bilingual texts with abundant pre­faces in some of them (31,3 - 69,38): Hadriani sententiae et epistulae, fa­bles of Aesop, the so called Dositheanum fragmentum de manumissione, 49

Hygini Genealogia, and an Iliad paraphrase. The L colloquia which, as I mentioned before, are a shorter version of S 1, follow abruptly after the last of these texts, the Iliad paraphrase, which lacks the beginning and the end. We find no traces of a collection of texts of this kind in groups B, C, and M. As I said, Goetz considered the L version original; Dionisotti alternatively suggested that it would be "a late gathering of originally separate material of this kind"; according to her, the dating in the Hyginus preface may have come from the colophon or title of an original work by Hyginus.SO Hadriani sententiae survive inS without the preface (387-390); a more complete ver­sion of Hygini Genealogia has also been combined with the glossaries of the Hygini Hermeneumata (H 72-94), as the excerpts in some manuscripts show.51

C, title p. 97.

4 7 C f. 18r; see Dionisotti, Schoolbook 93 and pl. Ill.

48 I cite the date as printed in CGL Ill 56,30-34: Ma~t~ro · Kat · anpro I una-rote I npo · y 0 tbrov ° C£1t-rePptrov I uytvou . Y£V£ . a'Aoytav I 1tUCtV . yvroC'tT\V . !-Le'teypa<pua -Maximo · etapro I consulibus I tertio id septeber I yginigenealogiam I omnibus notam descripsi. Note that ~-t£-reypa'lfa (cod. ~-t£-rc:ypa<pua) - descripsi here means "I copied", Dionisotti, Schoolbook 89. The absolute dative in the Greek is a common phenomenon in the consular dates of documents and indicates a translation from Latin. The only consuls that match are those of 207, Annius Maximus and Septimius A per.

49 The beginning of a Roman law primer of the same type as Gaius' Institutiones.

50 Dionisotti, Schoolbook 90.

51 See Dionisotti, Glossary Stories 327-30. These Hermeneumata also had the Coli.

114 Kalle Korhonen

The first problem here is that these five texts may well have existed in 207. The two datable texts in the collection, the Law fragment and Hadriani sententiae have both been dated - by their contents, not the consular dating - to the late 2nd century. 52 The dating of the other texts is more difficult. Babrius wrote, at the latest, in the 2nd century, as P. Oxy. 1249 shows.53 The other fables of Aesop and the Genealogia of Hyginus are almost undatable. 54 Nor is it easy to date the translations.

But let us turn to the prefaces of these texts. 55 According to Dionisot­ti, 56 the similarities in them are due to the formulaic nature of such prefaces. In my opinion, they rather seem to form a coherent whole. All the prefaces of the first three texts - leaving aside the first book Ill preface discussed ear­lier - emphasize that they help learning both Latin and Greek; the languages are always mentioned in this order.57 Another interesting feature of the pre­faces are the references to three-book material. A piece of the preface of book II comes up in a surprising place (47,58 - 48,7, preface of the law primer). As Dionisotti has noted, 58 the Hyginus preface also has a reference· to book 11 (capitula glossary): the author says that in the book that follows there will be many translated stories about gods and goddesses, whereas in

Harl. The Hyginus seems to have had no preface ( op. cit. 330). 52 The law primer: see A.M. Honore, RIDA, 3e ser. 12 (1965) 306-11, 323; H.L.W. Nelson, Uberlieferung, Autbau und Stil von Gai Institutiones ( 1981) 368-70. Hadriani sententiae: A. A. Schiller, in: Atti del secondo Congresso internazionale della Societa italiana di storia del Diritto II ( 1971) 720-24 (note that the linguistic analysis of the text announced by Schiller was never fully accomplished, cf. P. Stein, in Studies in Roman Law in Memory of A. Arthur Schiller (1986) xvii); on the text, see also F. Millar, The Emperor in the Roman World (1977) 532. 53 See also M.J. Luzzatto & A. La Penna (ed.), Babrii mythiambi Aesopei (1986) x-xi.

54 In my opinion, this Hyginus fragment should have been treated more thoroughly in the new Teubner edition by P.K. Marshall (Hygini Fabulae, 1993); the text is not even published in it, cf. Marshall, op. cit., p. x. For a more valid evaluation of the fragment, see L.D. Reynolds, in: Texts and Transmission (1983) 190. On the fables, see M. Npjgaard, La fable antique II (1967) 398-403. 55 The prefaces: Hadriani sententiae: 31,3-23; fables: 38,30- 41,5 (and 94,1 - 95,39 from a Paris mss. with some differences); the law primer: 47,58 - 48,45 (the first 8 lines seem to come from the preface of book II); Hygini Genealogia: 56,27 - 57 ,42. See also Dionisotti, Schoolbook 89 and Flammini, Prolegomeni 14, 18, 21-26. 56 Schoolbook 90. 57 L 31,20-21; 38,52-53; 48,19-24. 58 Dionisotti, Schoolbook 89.

On the Composition of the Hermeneumata Language Manuals 115

book II he has only listed their names (56,35-43).59 He also says that pic­tures illustrate them (56,47-50), as in the preface of Aesop (L 39,55). If the dating of the collection to 207 is correct, it would mean that the three-book form existed already, say, in the 2nd century. In any case, the three-book collections seem to antedate the prefaces of the L version. 60 But in the L version, the phrasebook proper would have disappeared.

There is another common feature which separates these prefaces from those of books I and II: the prefaces of Aesop, the law fragment and Hygi­nus mention ypaJ..LJ..La'tcov 'tEXVll, 'tEXVll ypaJ.lJ..La'ttKft or ypaJ.lJ..La'ttKo{ (38,43; 48,8-10, 56,51-52), and translations from Greek to Latin and vice versa (48,16-24). It is clear that we are dealing with grammarian's teaching. Now, what languages can be studied with the help of these texts? The version in the most important Leiden manuscript really seems to be suited for bilingual education: the Hadriani sententiae and the fragment of the primer of Roman law are suited for the study of Latin, but it would be natural to study Greek with the fables and the mythological texts. Dionisotti saw the whole Herme­

neumata material as a largely western phenomenon, suited for the simulta­neous education of both languages, 61 and it is an incontestable fact that Latin was studied by very few, if any children in Greek grammarians' schools in this period. 62 The school scenes in the eo lloquia are a similar case: Greek children learning Latin at school would be an oddity before Late Antiquity. If this collection of texts existed in bilingual form already in 207, the material would have had to belong to the W estem school.

However, the utility of the "law" texts is a problem. Would Latin­speaking children study Greek with the help of such texts? We should ex­plain this by saying that their purpose has been to teach how to translate laws and other legal documents into Greek, which was a normal practice in governing the eastern parts of the Empire.63 But the material seems ideal for the eastern schools of Late Antiquity, where such jurisprudential texts, also

59 The words £~t:nA£~aJ.lEV - explicuimus (actually "unfolded") must here mean "list", because the author emphasizes the distinction between book II and the one he is writing, in which tcr'topiat - historiae or £~nrftcrt:t~ - enarrationes can be found.

60 This was also the view of A.A. Schiller, op. cit. in n. 52, 719. 61 Dionisotti, Schoolbook 91; see also ead., Grammars & Dictionaries 29. 62 See, e.g., A. Bataille, in: Recherches de Papyrologie IV (1967) 162-68; H. Maehler, in Actes du XV e Congres International de Papyrologie II ( 1979) 18-19. 63 J. Kaimio, The Romans and the Greek Language (1979) 75-80, 109-10.

116 Kalle Korhonen

Gaius, who is fairly contemporary with the law primer in L, 64 were studied

intensively, to judge from the papyri. 65 Instead, the study of a foreign

language with the aid of bilingual Hyginus or fables seems natural for both

Roman and Greek children. In my opinion, it is possible that the collection

of texts has been compiled, translated and the prefaces written about a

century later than 207. Thus, we still have no terminus ante quem for the

three-book collections.

In the glossaries, elements datable to the Late Antiquity are few. 66 S 1,

the continuation of the L colloquia, has no criteria for dating. As to the of­

ficial or military terminology in L, the two glossaries (27 ,36 - 28,8 and 28,9-

23) contain only words that were already in use in the first century AD. This

is the case in the other versions as well: in the single capitula glossaries with

the title 1t£pt apxov'trov- de magistratibus or similar (L 28,9-23; M 182,22-

60; E 275-276; Mp 297,33 - 298, 15; S 362; C f. 30v-31 v), the bulk of

official terminology always belongs to the early imperial period; late antique

terms are very rare. Constantinian terminology is lacking in the glossaries. 67

In addition, we know that Greek -speaking persons were using bi­

lingual capitula glossaries already in the 2nd century to study Latin, as the

papyri show. P. Oxy. 2660 (1st I 2nd century) lists vegetables and fishes

with a heading between the lists.68 The order in the Hermeneumata is the

same. P. Oxy. 3315 (1st I 2nd c.) contains signs of the zodiac followed by

names of winds with their heading. 69 In this case, the order of the words is

64 See H.L.W. Nelson, op. cit. inn. 52, 364-70.

65 See R.A. Pack, The Greek and Latin Literary texts from Greco-Roman Egypt2 (1965) nos. 2953-2993.

66 Dux- ryyquov at 27,39 is a general term, cf. H.J. Mason, Greek Terms for Roman Institutions (1974) 146; mulomedicus in another glossary (25,56) seems a late word, but it is probably a gloss which has come to the text at a later stage. A certain later addition is salomonis at 9,38, probably a reference to Salomo Ill, abbot of St. Gall (fl. ea. 900), compare Glossarium Salomonis. The problem here is that words are hard to date, and when glossaries are copied, words can be changed.

67 E.g. comes appears only once in CGL Ill, namely in the alphabetical dictionary of M, at 159 ,38; it is translated by cruvo8o{nopoc;.

68 It has been republished by Kramer, Glossaria 63-64 no. 6. Parts of the same two glossaries are also in the later P. Oxy. 2660a (3rd cen.; Kramer, Glossaria 67 no. 7). All the papyrus datings used here are by the editors. My thanks are due to Jaakko Frosen for help in evaluating the correctness of the datings.

69 Krarner, Glossaria 69 no. 8.

On the Composition of the Hermeneumata Language Manuals 117

not the same in any of the Hermeneumata versions. I also mention P. Mich.

inv. 2458 (2nd I 3rd c.) with the end of a list of gods, the beginning of an­

other with goddesses, and the heading of the second list in between. 70 It is

like a fragment from the beginning of a capitula glossary. Fragments of sin­

gle capitula glossaries have survived in P. Laur. inv. III/418 (late 2nd c.),71

and P. Lund 5 (2nd c.).72 In all of these five papyri, the Latin part has been

written with Greek letters. It has long since been noted that these papyri do

not seem to belong to the school sphere, if this means "children's school",

but have instead been written by skilled writers. Thus they are considered as

predecessors of the topical dictionaries of the modem travellers' vocabular­

ies, written for adults. 73 This matter needs further study, as it is difficult to

say who would have needed such vocabularies. Even if the word order is in

no case so similar between the papyri and Hermeneumata that we could say

that a papyrus belongs to a certain version,74 it is certain that one source

material of the Hermeneumata were glossaries similar to these. We do not

know if these glossaries belonged to or where copied from the three-book

collections.

The prose parts in the other Hermeneumata versions have elements

that cannot be earlier than the 3rd century, but their use may have continued

for some time. The expression domini mei imperatores in Coll. Harl. 9 and

the name Aurelius (Coll. Harl. 19) date the full Coll. Harl. to the third or

fourth century.75 In M, the key term is praeses provinciae (8t£nrov ti,v enapx{av, M 4 ), which dates the text to probably the late 3rd century or

70 N.E. Priest, ZPE 27 (1977) 193-200; Kramer, Glossaria 79-81 no. 12.

71 R. Pintaudi, ZPE 27 (1977) 115-117; J. Rea, ZPE 29 (1978) 240; Kramer, Glossaria 61-62 no. 5.

72 Also published by Kramer, Glossaria 71 no. 9. A later capitula glossary fragment is in P. Vindob. L 150 (5th c.; J. Kramer, Tyche 5, 1990, 37-39).

73 W. Brashear, in: Proceedings of the XVI International Congress of Papyrology (1981) 33-34; Kramer, Glossaria 10. See also A. Bataille, op. cit. inn. 62, 161-69; J. Kramer, in: Atti del XVII Congresso internazionale di papirologia (1984) Ill 1379-80, 1384.

74 The nearest affinities are provided by the late P. Vindob. L 150 mentioned inn. 72. But glossaries, when copied, tend to change much more than prose texts (essential points about the transmission of glossaries have been made by A.C. Dionisotti, in the new volume Les manuscrits des lexiques et glossaires de l'antiquite tardive a la fin du moyen age (1996) 205-25).

75 See K.J. Neumann, RE V (1905) 1305-09; G. Lugli, in: Dizionario epigrafico delle antichita romane II.3 (1922) 1954-55. On Coli. Harl. 9, see above, p. 106.

118 Kalle Korhonen

later.76 However, in the passage M 5 cited above (p. 105), in the Latin version sestertii are mentioned (M 5 = 212,54 = E 228,50), but not in the Greek. The use of sestertii ceased in the Diocletianic period; 77 here the Greek term has been changed to correspond to reality, but not the Latin one. The prices elsewhere in the colloquia do not help us much; but at least Constantinian solidi are not used as currency. It is clear that money terms and official terminology may have been updated in language manuals that were meant for the study of everyday language. The context of the colloquia in Mp is Rome; it is harder to date because it has no official terminology. The balneum Tigillinum (Mp 14) gives it a terminus a quo, the Neronian period. 78 C, on the other hand, has more clues for dating; Dionisotti dates it to the late 3rd or more probably 4th century.79 Dionisotti has plausibly suggested that we have one testimony of such manuals in the literature, Ausonius' partly preserved poem Ephemeris, id est totius diei negotium. 80 This indicates that the colloquia were used in the 4th-century West; Ausonius may have used them at school. The original capitula glossaries and the texts show no certain signs of Christianity, 81 but the phrase book material was still in use in 5th-6th century Egypt, as the two papyri mentioned above (p. 104) show. Together, these instances bear witness to a long period of use of the Hermeneumata material.

The traditional interpretation has been that the H ermeneumata were created for Greeks who wanted to learn Latin; later, they were also used by Romans to learn Greek. 82 The scholars have been puzzled by the fact that the prefaces seldom speak of only one language, and normally both the languages are named. As a consequence, it has also been proposed that they

76 See W. EnBlin, RE Suppl. VIII (1956) 602. 77 K. Regling, RE IIA (1923) 1882. 78 The name probably refers to Ofonius Tigellinus (PIR 2 0 91 ); the balneum is attested in other sources, too, see E. Rodrfguez Almeida, in: Lexicon topographicum urbis Romae I (a cura diE. M. Steinby) (1993) 165. 79 Dionisotti, Schoolbook 122-23; see also A. Giardina, RFIC 113 (1985) 316-20. The most clearly datable part in the C text is, however, the schoolbook sequence C 70-77 (see above, p. 1 07) .. 80 Dionisotti, Schoolbook 123-25. Cf. R.P.H. Green, The Works of Ausonius (1991) 246. 81 Dionisotti, Schoolbook 91. 82 Goetz, CGL I (1923) 18; J. Tolkiehn, RE XII (1925) 2468; Marrou, Histoire (seen. 1) II 59, n. 20 p. 194.

On the Composition of the Hermeneumata Language Manuals 119

were designed for the needs of both groups. 83 Dionisotti's criticism against the first theory is justified. I think that the problem needs to be clarified fur­ther, but I consider possible that the authors of the phrase book and glossary material really would have had in mind the speakers of both languages.

5. Conclusions

I propose that the Hermeneumata we have in the manuscripts have been put together from two different kinds of material: 1) everyday conver­sations an~ vocabularies for practical use by adults; 2) school texts more adapted for translation practice. I think this is the only way to explain the differences in the characters and in the manner of narration. The distinction between the glossaries based on the studies of grammarians and on the needs of everyday communication is useful, but the Hermeneumata we have now contain both types of material. The prefaces of the alphabetical and ca­

pitula glossaries we have seem to belong to the practical material. The con­text of this material, which has already been known from the papyri, needs to be studied further.

I have also discussed the problematic Leiden collection. In my opin­ion, the prefaces of the bilingual texts form a coherent whole. They indicate that the compiler of the bilingual version has used an existing three-book Hermeneumata collection. It would seem that the collection belongs to the study of Latin as a foreign language in Late Antiquity. However, the rest of the Hermeneumata material shows signs of a long period of use. It is also probable that the compilers of the three-book H ermeneumata collections had in mind both the learners of Greek and the learners of Latin.

University of Helsinki

83 J. Kaimio, The Romans and the Greek Language (1979) 203-04; see also Marrou, Histoire II n. 20 p. 194. It is improbable that they would have been written for an audience that knew neither Greek nor Latin.